This article originally appeared on Inc Magazine.

Do it as soon as possible, while preparing both your company and your people

As more organizations adopt AI into everyday operations, the gap between those who choose to innovate with this technology and those who do not is only going to widen. The movement to adopt AI will not be limited to the industries that need these skills today, like renewables or chip manufacturers, it will include nearly every industry in existence.

The writing is on the wall: three out of four C-suite executives believe that if they don’t scale AI in the next five years, they risk going out of business entirely. Organizations need to incorporate AI and natural language processing into their businesses now to be viable and competitive.

The good news is that many are already doing so, but these initiatives tend to be one-offs or experimentation only. And while that’s valuable, how can companies go from this point to successfully scaling AI across the organization?Start with a mindset shift

There was a major disruption in “business as usual” when the internet came into play in the late 1990s. There was also a mindset shift that took place — tasks could be done in a digital world faster and more efficiently than in a manual world.With AI, understanding where the business wants to be and what skills will be needed to get there will require organizations to take a step back and think differently. Focusing on understanding employee skills rather than on job descriptions is a departure from “business as usual.” Identifying declining and rising skills enables organizations to create the paths for the workforce to build those rising skills they’ll need for the future of work.

AI is poised to change the way people work, not necessarily replace people who do outdated work.

Just as with any other technological innovation, whether AI improves jobs or eliminates them will depend on how it’s deployed. If the mindset is to use AI to create value for the customer and improve the jobs of employees, then AI can be a great asset.Many organizations are already doing this now. For example, major retailer Dick’s Sporting Goods adjusted its talent strategy by embracing new skills to serve customers better and create internal opportunities for employees. Using AI, the company is well on its way to building the skills of its nearly 50,000 employees to deliver the kind of experience its customers demand.

Identify business challenges with a proof of conceptThis is where most organizations get stuck. They conduct AI pilots or experiments but fail to achieve a return on investment, which hinders their ability to scale. The main issue here is putting the cart before the horse.

The question should be: What business challenges need to be solved, and how can AI help solve them?

Tie the pilot or experiment to a business challenge, need, or strategy and brainstorm ideal metrics. This way it will be evident whether the pilot delivers the expected ROI. Also, siloed initiatives often stay inside one department or team. It’s good to think holistically by engaging other teams, especially the C-suite.When the C-suite can see overarching analytics that show AI can help the organization meet business objectives, increase revenue, and keep customers happy, that’s when there’s buy-in.

The other issues with scaling AI beyond pilots are unrealistic time frames and under-investment, which yields low returns. According to an Accenture study, a “lack of budget” was at the bottom of the list of pilot challenges. The key to getting the most value out of AI is scale.Scale strategically, focusing on the big picture

It’s not just about rolling AI out across the entire organization as quickly as possible. Scaling requires strategy. Companies that strategically scale AI report nearly three times the return from AI investments compared to companies pursuing siloed proof of concept.

Focus on “big picture” problems using advanced analytics to guide the way. It’s easy to get bogged down in the details of the data, so keep a high-level view.Establish realistic expectations in terms of time to scale. It’s better to go slow and steady, making sure business objectives are being met, than it is to speed up with no proof of ROI.

To ensure that AI is scaled responsibly, it’s a smart move to establish an employee resource group. This group can not only act as a champion of AI across the organization but also create accountability and align efforts. They can also set how the organization will measure success from AI.Make continuous real-time insights the end goal

The end goal of scaling AI is to create a culture of continuous real-time insights that drive business decisions, so the organization can innovate to consistently improve customer satisfaction.

The best way to reach this end goal is with a clear vision across the organization that includes communication, accountability, and metrics, so everyone knows and trusts in the business practices. The vision should also include ongoing training for employees, so they understand how AI applies to their roles within the organization.Above all, incorporating AI into any organization demands a shift in mindset. If the prevailing attitude is that people are a cost to be minimized, then AI will simply be a substitute for labor. However, by prioritizing skill development and fostering a culture that views AI as a tool to enhance, rather than replace, human potential, organizations can change the way people work for a more positive impact that benefits both the organization and its people.